Geodatabase (Esri)

The origin of the geodatabase was in the mid-1990s during the emergence of the first spatial databases. One early approach to integrating relational databases and GIS was the use of server middleware, a third-party program that stores the spatial data in database tables in a custom format, and translates it dynamically into a logical model that can be understood by the client software. In 1996, Esri purchased an early middleware product called Spatial DataBase Engine and rebranded it ArcSDE. Initially, ArcSDE stored and delivered simple vector datasets that looked very similar to shapefiles, but the need for a more robust data model emerged as Esri's Shapefile format became a de facto standard for vector spatial data, even as its shortcomings limited its use in enterprise applications. At the same time, the Arc/INFO coverage format was becoming obsolete after 20 years, unable to handle growing expectations of GIS users. Another motivating factor was that even though several relational database vendors were introducing their own spatial extensions (with the notable exception of Esri's preferred Microsoft SQL Server), their structures and interfaces varied and Esri wanted its users to see all spatial data in the same apparent structure regardless of how it was stored internally. At the end of 1999, Esri introduced the Geodatabase model as the native format used in its new ArcGIS software (branded Version 8.0 to maintain continuity with Arc/INFO). Initially, it could be implemented as a multiuser geodatabase in ArcSDE on a server or the personal geodatabase locally. Support for topology rules, linear referencing, and survey data were added in 2003 (with ArcGIS 8.3). Network data was added to the geodatabase in 2005 (ArcGIS 9.1), and vector terrain (TIN, LIDAR) in 2006 (ArcGIS 9.2). Also at the 9.2 release, ArcSDE was subsumed into ArcGIS Server and the multiuser database format was rebranded the enterprise geodatabase. Due to shortcomings in the personal geodatabase format (especially file size limitations in Microsoft Access), Esri developed a more robust custom file format, released in 2006 (ArcGIS 9.2) as the file geodatabase. Eventually, the middleware components for reading and writing the geodatabase spatial database structure were incorporated into ArcGIS desktop, eliminating the need for ArcSDE to be running on the server end. The most recent addition has been the mobile geodatabase format in 2020 (ArcGIS Pro 2.7), which uses SQLite as the backend to store the entire geodatabase as a single file. This replaces the personal geodatabase, which is no longer supported. ==Applications==

Geodatabases, being a common format for GIS datasets, have applications anywhere GIS are widely employed. These applications are so basic oftentimes researchers do not mention their use in studies. There are several fields where their use is extensively documented, including public health, crime analysis, and resource management. Epidemiology Since John Snow famously identified the source of a cholera outbreak, spatial data has been central to epidemiology and public health. In recent years, information that is relevant to public health has increased exponentially. Leveraged correctly, this data can allow for a rapid response to emerging diseases. To accomplish this, geodatabases are employed extensively to organize data and allow for the identification of space-time patterns. They were used extensively to organize data related to West Nile virus epidemics, and the COVID-19 pandemic. This use includes analyzing misinformation, and the infodemic, surrounding COVID-19. Resources management Geospatial data around resource management plays is extremely complex. Factors such as the forest, water, and mineral resources being managed are obvious; however, governance and socioeconomic factors also play a large role. It is common practice to employ geodatabases to manage these diverse datasets. TIGER files In 1995 The United States Census Bureau made the Topologically Integrated Geographic Encoding and Referencing, or TIGER, Mapping Service available to the public, facilitating desktop and Web GIS by hosting US boundary data. This data availability, facilitated through the internet, silently revolutionized cartography by providing the world with authoritative boundary files, for free. Today, these files, which contain up-to-date boundaries for the United States states, counties, and more, are provided to the public in prepackaged geodatabases. ==Logical Model==

To the user, a geodatabase looks like a collection of datasets, including some containing geographic data and some auxiliary elements that add functionality to the data. This user view is identical, regardless of how the geodatabase is stored (although enterprise geodatabases add some functions). Datasets Datasets contain geographic data. A geodatabase can contain spatially referenced data in vector or raster formats, or non-spatially referenced data in tabular format. Each dataset contains information about any number of individual items, but typically all of the items in a dataset are of the same theme (e.g., temperature measurements, roads in a city) and have the same set of properties. ; Table : A traditional relational database table without any spatial information. Called a "business table" in early versions of ArcGIS. Each vertex in a line can be three-dimensional (x,y,z) and can also store a measurement value (e.g. a highway milepost value). Lastly, line segments can be curved as circular arcs or Bézier curves in addition to the traditional straight lines. Geodatabase rasters using internal tiling and pyramid structures to improve drawing and analysis performance, especially for very large grids. Auxiliary elements A number of elements can be included that are generally dependent on one or more datasets, adding functionality such as quality control. Some of these are called controller datasets ; Feature dataset : A collection of several feature classes within the geodatabase, roughly similar to a folder in a file system. All of the feature classes in a feature dataset share the same spatial reference system. These facilitate data entry (the user can pick a value from a list rather than trying to remember the exact text) and quality control (reducing invalid values) ; Subtype : Partitions a table or feature class into subcategories, with common attributes and other behaviors. ; Topology : A list of rules for valid topological relationships between features in one or more feature classes (e.g., "county polygons cannot overlap," "state polygons must align with county polygons"). ; Terrain dataset (formerly TIN dataset) : A triangulated irregular network constructed from feature classes containing three-dimensional points and lines. ; Network dataset, Utility network : Two different approaches to constructing a topologically-connected network from feature classes representing linear segments (and often point junctions). ; Linear referencing system (LRS) : An augmentation of a line feature class that enables attributes to referenced to points or segments within each line rather than attached to the line as a whole. Examples include storing accident locations or real-time traffic counts along a road. ; Mosaic dataset : A virtual composite raster grid composed of images that are stored as separate raster files. This is stored as a polygon feature class with a row for each image including image properties such as image filenames and georeferencing information, and the shape representing the desired display extent of the image, enabling seamless composition of overlapping images. ==Implementations==

Since its first introduction in 1999, the geodatabase has been available on a number of platforms to meet various project needs. ;Enterprise Geodatabase (formerly Multiuser Geodatabase) :The data is stored in a third-party relational database management system on a server. This allows for the implementation of the geodatabase within various distributed GIS architectures, including Internet and Web GIS through the use of ArcIMS (Arc Internet Mapping Service). Originally, only commercial RDBMS software was supported: Microsoft SQL Server, Oracle, Informix, or IBM Db2. Eventually, support for PostgreSQL and SAP HANA were added and Informix was discontinued. The spatial data in a feature class can either be stored using the geometry datatype provided by the RDBMS native spatial extension (Oracle Spatial, PostGIS, etc.) or Esri's proprietary ST_GEOMETRY format, all of which are based on the Open Geospatial Consortium Simple Features specification, but with different encoding structures. • GDB_Items: a "table of contents" for all of the elements of the geodatabase as the user will see them, pointing to the corresponding physical tables • GDB_ItemTypes: the type of dataset of each table (table, feature class, etc.) • GDB_ItemRelationships: information about groupings of tables, such as feature datasets • GDB_ItemRelationshipTypes: lookup table of types of item relationships • GDB_DBTune: general parameters for the geodatabase • GDB_SpatialRefs: a list of the spatial reference systems used in the datasets • GDB_SystemCatalog: a list of all tables, including data and system tables ;Personal Geodatabase :One of the original options, storing data in a single Microsoft Access file (.mdb), intended for local storage for a single user without needing a server. and Esri provides a read/write library for use in other software. It is not a single file, but a collection of files (roughly one for each data or system table in the relational database geodatabase) collected in a folder with a .gdb extension. The following file types are included: Initially released with limited functionality, it now supports all of the geodatabase elements. Despite the common database, this is very different from a Geopackage model. ==Application support==

The ability to read and write geodatabase format is not limited to Esri products; other software are also able to read & write this format, including: • ENVI – Remote sensing software owned by Harris Geospatial • ERDAS IMAGINE – Remote sensing company owned by Hexagon AB • • ==See also==